线性模型中一个特征矩阵的研究

Research of a characteristic matrix in linear model

目的 研究了线性模型大样本情形下S-1n的极限特征。方法 利用投影分析方法并结合其统计背景分析了S-1n的结构形式。结果 在一定条件下证明了当n→∞时S-1n主对角元本身具有并且同时从整体上决定了其他各元的下降特征。结论 与前人结果相比,提供的结果把S-1n的极限特征进一步深化,在线性模型估计的相合理论中应用更方便。

AimTo discuss the limiting character of S^(-1)_n in the case of large samples in linearmodel.MethodsThe structure of S^(-1)_n is got by the method of projection analysis under its statistical surrounding.ResultsIt is shown that main diagonal elements of S^(-1)_n is descending themselves when n→∞,which makes the other elements of S^(-1)_n descending on the whole at the same time.ConclusionCompared with the corresponding results of the others,the conclusion in this paper discussed the limiting character of S^(-1)_n more deeply, which will lead to a more convenient appliance in the consistent theory of linear models.